Buprenorphine treatment of opioid dependence: analysis of individual patient data

BackgroundThe efficacy and safety of buprenorphine alone and in combination with naloxone for treatment of opioid dependence were evaluated in Federally-sponsored randomized clinical trials. Meta-analysis of pooled individual participant data provides an opportunity to identify multiple predictors of buprenorphine treatment outcome. MethodsWe selected six buprenorphine efficacy and safety trials from NIDAs Data Share database for analysis. Treatment, sociodemographic, and drug use history variable domains were systematically harmonized and included in analysis. After exclusions, 3,022 participants randomized or enrolled in buprenorphine treatment for opioid dependence (mean (SD) age 36.1 (9.8) years, 33% female, 66% White, 16% Hispanic, 14% Black), were analyzed using a generalized linear mixed model with time-weighted treatment variables and participant covariates. We defined positive urinalysis or self-reported lapse as the primary outcome. ResultsFour treatment variables were significantly associated (p < 0.001) with lapse. Time-weighted dose and time-weighted adaptive dose had greater estimated effects than time-in-trial and time-weighted clinic visit. All treatment variables were novel predictors of lapse. ConclusionsIn a large cohort of trial participants treated with buprenorphine and behavioral counseling for opioid dependence, we identified and ranked four novel treatment factors reflecting components of buprenorphine dose, clinical provider engagement and patient engagement. Additional research to explore the effects of pharmacologic and non- pharmacologic treatment factors, and to explore relations with provider and patient factors will help our understanding of buprenorphine treatment outcomes. Continued analyses of publicly available data will extend discovery and support development of personalized opioid use disorder treatments. Highlights (3 to 5 bullet point max 85 characters each including spaces)O_LITreatment and participant variables were harmonized in six buprenorphine trials C_LIO_LITime-weighted treatment variables were used in a random effects mixed model of lapse C_LIO_LIBuprenorphine dose and three clinical interactions were protective against lapse C_LIO_LISupport of protective treatment factors may improve buprenorphine treatment success C_LI